Experimental and chemical kinetic study of the pyrolysis of trifluoroethane and the reaction of trifluoromethane with methane

Article abstract of DOI:10.1016/j.jfluchem.2010.03.012

A detailed reaction mechanism is developed and used to model experimental data on the pyrolysis of CHF₃ and the non-oxidative gas-phase reaction of CHF₃ with CH₄ in an alumina tube reactor at temperatures between 873 and 1173 K and at atmospheric pressure. It was found that CHF₃ can be converted into C₂F₄ during pyrolysis and CH₂=CF₂ via reaction with CH₄. Other products generated include C₃F₆, CH ₂F₂, C₂H₃F, C₂HF ₃, C₂H₆, C₂H₂ and CHF₂CHF₂. The rate of CHF₃ decomposition can be expressed as 5.2×10¹³[s^-1]e ^-295[kJmol^-1]/RT. During the pyrolysis of CHF₃ and in the reaction of CHF₃ with CH₄, the initial steps in the reaction involve the decomposition of CHF₃ and subsequent formation of CF₂ difluorocarbene radical and HF. It is proposed that CH₄ is activated by a series of chain reactions, initiated by H radicals. The NIST HFC and GRI-Mech mechanisms, with minor modifications, are able to obtain satisfactory agreement between modelling results and experimental data. With these modelling analyses, the reactions leading to the formation of major and minor products are fully elucidated.

Full text of DOI:10.1016/j.jfluchem.2010.03.012

Journal of Fluorine Chemistry 131 (2010) 751–760
Contents lists available at ScienceDirect
Journal of Fluorine Chemistry
journal homepage: www.elsevier.com/locate/fluor
Experimental and chemical kinetic study of the pyrolysis of trifluoroethane
and the reaction of trifluoromethane with methane
a
a
b
Wenfeng Han a, Eric M. Kennedy a, , Sazal K. Kundu , John C. Mackie , Adesoji A. Adesina ,
*
Bogdan Z. Dlugogorski a
a Process Safety and Environment Protection Research Group, School of Engineering, The University of Newcastle, Callaghan, NSW 2308, Australia
b Reactor Engineering and Technology Group, School of Chemical Sciences and Engineering, University of New South Wales, Sydney, NSW 2052, Australia
A R T I C L E I N F O
A B S T R A C T
Article history:
A detailed reaction mechanism is developed and used to model experimental data on the pyrolysis of
CHF3 and the non-oxidative gas-phase reaction of CHF3 with CH4 in an alumina tube reactor at
temperatures between 873 and 1173 K and at atmospheric pressure. It was found that CHF3 can be
converted into C2F4 during pyrolysis and CH255CF2 via reaction with CH4. Other products generated
Received 11 February 2010
Received in revised form 15 March 2010
Accepted 19 March 2010
Available online 27 March 2010
include C3F6, CH2F2, C2H3F, C2HF3, C2H6, C2H2 and CHF2CHF2. The rate of CHF3 decomposition can be
ꢁ1
expressed as 5:2 ꢀ 1013 ½sꢁ1ꢂ eꢁ295½kJ mol ꢂ=RT . During the pyrolysis of CHF3 and in the reaction of CHF3
Keywords:
with CH4, the initial steps in the reaction involve the decomposition of CHF3 and subsequent formation of
CF2 difluorocarbene radical and HF. It is proposed that CH4 is activated by a series of chain reactions,
initiated by H radicals. The NIST HFC and GRI-Mech mechanisms, with minor modifications, are able to
obtain satisfactory agreement between modelling results and experimental data. With these modelling
analyses, the reactions leading to the formation of major and minor products are fully elucidated.
ß 2010 Elsevier B.V. All rights reserved.
HFC 23
CHF3
Greenhouse gas
Vinylidene fluoride
Hydrofluorocarbon
Kinetic modelling
1. Introduction
as outlined in United Nations Framework Convention on Climate
Change (UNFCCC) [4]. In this process, steam, O2 and natural gas are
Global production and use of chlorofluorocarbons (CFCs) and
halons has decreased significantly as a result of the phase out
schedules of the 1987 Montreal Protocol and its subsequent
amendments and adjustments [1]. The consumption and emission
of hydrofluorocarbons (HFCs) are projected to increase substantially
in the coming decades in response to regulation and replacement
of ozone depleting gases under the Montreal Protocol [2].
Trifluoromethane (CHF3, HFC-23, fluoroform, FE-13, R23) is an
unintentional by-product generated during the manufacture of
CHClF2 (HCFC-22), and has limited application as a refrigerant or as
a raw material for other products. It has a global warming potential
11,700 times greater than carbon dioxide and an atmospheric
lifetime of 264 years. It is reported that the concentration of CHF3 is
steadily increasing in the atmosphere since at least 1978 at a
present rate of increase of 5% per year [3].
used as reactants or fuels for the decomposition of CHF3 at
temperatures as high as 1473 K. It is also reported that phosphates
and ZrO2–SO4 are active and stable catalysts for the destruction of
CHF3 in the presence of O2 and steam at relatively low temperatures
[5,6]. One of the problems with oxidation is that the fluorine has to
be removed as HF from the exhaust gas stream and then recycled or
disposed of as a fluoride salt. More recently, Moon et al. investigated
the pyrolysis of a mixture of CHF3 and tetrafluoroethylene (TFE) as
a potential route for the production of hexafluoropropene (C3F6)
and found that the addition of CHF3 can significantly increase the
yield of C3F6 compared to the pyrolysis of TFE alone [7].
In a previous work, we reported that CHF3 can be converted to
vinyl difluoride (VDF, CH255CF2), a highly valuable feedstock,
through reaction with CH4, although the conversion and yield of
target product were relatively low [8,9]. VDF is widely used as a
fluoroelastomer which is a key monomer for the synthesis of a
variety of products, most notably poly-(vinylidene fluoride)
(PVDF), Viton (produced by Dupont Corporation), KEL-F (produced
by 3M) and Aflas (produced by Asahi Glass Co. Ltd.). Components
fabricated from fluoroelastomers enhance reliability, safety, and
environmental friendliness in such areas as automotive and air
transportation, the chemical processing industries, and in power
generation [10]. A wide range of fluoroelastomer products have
been developed to meet performance requirements in many
Thus, the development of suitable methods for the treatment of
CHF3 is of great practical significance. However, limited research has
focused on the development of treatment processes specifically for
CHF3, especially in comparison with the intensive research on
technologies for destruction of halons or CFCs. One method, thermal
oxidation, is a demonstrated technology for the destruction of CHF3
* Corresponding author. Tel.: +61 2 4985 4422; fax: +61 2 4921 6893.
E-mail address: eric.kennedy@newcastle.edu.au (E.M. Kennedy).
0022-1139/$ – see front matter ß 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2010.03.012

752
W. Han et al. / Journal of Fluorine Chemistry 131 (2010) 751–760
hostile environments, and to attain fabrication characteristics
comparable to other elastomers. More recently, VDF was found to
be an excellent source for the preparation of paints for high
performance external architectural coatings [11].
4. Results and discussion
4.1. Thermal pyrolysis of CHF3
The motivation of this study is to compare the thermal
decomposition of CHF3 alone and in the presence of CH4. The
conversion approach adopted in this work is distinctly different to
conventional destruction processes, as we aim not to destroy
fluoroform but rather to transform it to a useful product. Through
the combination of experimental work and kinetic modelling
simulation presented in this manuscript, the elementary reactions
dominating CHF3 decomposition, CH4 activation and the formation
of products can be fully elucidated.
In order to investigate the pyrolysis of CHF3, the conversion of
CHF3 at temperatures from 973 to 1073 K and at 1.01 bar versus
residence time was studied. Under these conditions, the conver-
sion level of CHF3 in a 10% CHF3-90% N2 pyrolysis mixture is
generally below 10%. Hence, reaction rate of this diluted mixture
can be approximated as
ꢁrA ¼ kCAn
(1)
which integrates to
1
2. Experimental
ln
¼ kt
(2)
1 ꢁ X
The pyrolysis of CHF3 and reaction of CHF3 with CH4 was carried
out in a tubular alumina reactor with an inner diameter of 7.0 mm.
This experimental facility has been described in detail elsewhere
[12,13]. Briefly, the apparatus consists of a tubular high purity
(99.99%) alumina reactor. Carbon containing products were
identified by a GC/MS (Shimadzu QP5000) equipped with an AT-
Q column, and quantified with a micro-GC (Varian CP-2003)
equipped with molecular sieve 5A and PoraPLOT Q columns.
Relative molar response (RMR) factors of hydrocarbons and
halogenated compounds for TCD detection were experimentally
obtained from standard gas mixtures where possible. The RMR of
species where standard gas mixture were not available were
estimated from published correlations [14]. HF formed during the
reaction was trapped with 0.1 M NaOH solution, and concentra-
tions were determined by an ion chromatograph (IC) (Dionex-100)
equipped with an IonPAS14A column (4 mm ꢀ 250 mm).
for a first-order (n = 1) mechanism, or
X
¼ CAokt
(3)
1 ꢁ X
for a second-order (n = 2) mechanism.
Where k is the reaction rate constant ((mol cm )
ꢁ3 1ꢁn sꢁ1), CA is
the concentration of CHF3 (mol cmꢁ3), rA is the rate of the reaction
(mol cmꢁ3 sꢁ1), X is the conversion of CHF3 and t reaction time.
We evaluate the pyrolysis kinetics of CHF3, assuming ideal plug-
flow conditions and a constant density system. The first-order and
second-order equations were used to fit the experimental data and
first-order assumption best matches the experimental data, as
shown in Fig. 1. The apparent rate constants for CHF3 decomposi-
tion in the temperature range from 973 to 1073 K, based on a least-
squares fitting of the experimental rate constants with an
Arrhenius expression is shown in Fig. 1. The rate constant
expression for the first-order reaction is given by;
The gases and solid reactants used in this study include CHF3
(>98%, Coregases), CH4 (99.99%, Linde) and AlF3 (Sigma, >99%). In
reactions which involved AlF3, 0.2 g aluminium fluoride was
charged into the uniform zone of the reactor, and held in place by
alumina chips (99.99%). Prior to reaction, aluminium fluoride was
dried in situ in a nitrogen atmosphere (99.999%, Linde) for 2 h at
673 K and 1.5 h at 1073 K. Feed gases, diluted in nitrogen (99.999%,
Linde), were introduced to the reaction zone.
ꢁ1
k ¼ 5:2 ꢀ 1013 ½sꢁ1ꢂ eꢁ295ꢃ46 ½kJ mol
(4)
ꢂ=RT
The pyrolysis of CHF3 was first studied using shockwave
techniques in the temperature range of 1200–1600 K [18,19]. Since
then, various results have been reported based on shock wave
experiments [20–22] or RRKM theoretical calculations [23]. These
derived rate expressions are summarized in Table 1, along with the
pressures and temperatures under which the data were obtained.
It is generally agreed that the initial step in the decomposition of
CHF3 is the dehydrofluorination and formation of CF2 species. We
will discuss the reactivity of singlet and triplet states of CF2 in more
detailed in Section 4.2.2.
As shown in Table 1, the values of A and Ea obtained from our
experimental data are close to the results of Placzek et al.’s RRKM
calculation [23] and Politanskii et al.’s thermal pyrolysis experi-
ments [24]. However, these rate constants are significantly lower
than those reported by Tschuikow-Roux et al. [18,19]. Biordi et al.
[25] studied the flame structure of bromotrifluoromethane-
inhibited methane flames and found that the rate expressions
given by Tschuikow-Roux et al. were too large to be consistent with
low-pressure flame data. Using similar experimental techniques,
an even lower activation energy for CHF3 decomposition was
suggested by Modica et al. behind incident and reflected shock
waves over a temperature range from 1600 to 2200 K [20]. One
possible reason for this discrepancy, as acknowledged by
Tschuikow-Roux et al. [18], is the difficulty to appreciably vary
the reaction dwell time while maintaining constant reaction
temperature and pressure conditions in the single-pulse shock
tube. Another reason for the discrepancy is that the decomposition
of CHF3 may be pressure-dependent and lie in the fall-off region
near the second-order limit [19].
3. Chemical kinetic modelling
The pyrolysis of CHF3 and reaction of CHF3 with CH4 have been
modelled using the commercial software package Cosilab [15].
During simulations, the steady state material balance for each
species was performed. As all experiments were conducted under
essentially isothermal conditions, energy balances were not
undertaken. Successive grids tolerance for species profiles were
set to 0.001(GRAD parameter) for species concentration and to
0.01 for the concentration gradients (CURV parameter). The final
grids contained 150 mesh points. The kinetic mechanism and
thermodynamic database used for reaction of fluorinated species
was the NIST HFC mechanism [16] with oxygen chemistry deleted
since there is no oxygen in the reacting systems. Gas Research
Institute GRI-Mech [17] was used for the pyrolysis of CH4, again
with oxygen-containing species removed. For reaction of CHF3
with CH4, NIST HFC mechanism and GRI-Mech were combined and
called GRI-NIST mechanism in this study.
Generally, a reasonable agreement of predictions of NIST HFC,
GRI-Mech and GRI-NIST mechanisms for CHF3 pyrolysis, CH4
pyrolysis and reaction of CHF3 with CH4 with experimental data
was obtained. However, deviations between experimental data
and modelling predictions were found in some cases. Modifica-
tions to the mechanisms are suggested, and discussed.

W. Han et al. / Journal of Fluorine Chemistry 131 (2010) 751–760
753
As C2F4 was observed as the major product in their experi-
ments, Tschuikow-Roux et al. suggested that the vibrationally
excited C2F5H* disproportionates rapidly into C2F4 plus HF.
However, we suggest that reactions (R3a) and (R3b) play a
relatively minor role under our experimental conditions, as there
does not appear to be any reported observations of CF2 insertion
into CHF3, and no evidence was found for CF2 inserting into C–H of
CH4 or C2H4 over the temperature range of 295–873 K [26].
According to our quantum chemical calculations, it is found that
CF2 can insert into C–H of CH4 with an activation energy as high as
163 kJ molꢁ1. The exponential factor was estimated to be only
1.5 ꢀ 1011 cm3 molꢁ1 sꢁ1. As argued in the reference [21], if
reaction (R3a) is responsible for the consumption of CHF3, we
would anticipate the formation of CF4, CF2H2, C2F6, C2F4H2 and
C2F5H, as a result of the following reactions:
CHF3 þ CF2 ! CF3 þ CHF2
CHF3 þ CF3 ! CF4 þ CHF2
CHF3 þ CHF2 ! CF3 þ CH2F2
(R4)
(R5)
(R6)
CF3 þ CF3 ! C2F6
(R7)
(R8)
(R9)
CHF2 þ CHF2 ! C2H2F4
CHF2 þ CF3 ! C2F5H þ M
Furthermore, thermal decomposition of CHF3 can be simulated
very well when reaction (R3) is excluded [7,8,21]. Further
investigations were undertaken to elucidate the mechanism of
CHF3 pyrolysis reaction and to estimate their associated kinetics
parameters.
A
feed mixture of N2 and CHF3, in which CHF3 was
approximately 10% by volume, was investigated at temperature
range of 873–1173 K and at atmospheric pressure. The residence
time was maintained at around 0.5 s by adjusting the volume of the
reaction zone. The conversion of CHF3 and the rate of product
formation as a function of temperature are shown in Fig. 2. It is
seen that CHF3 pyrolysis commences around at 973 K at residence
time of 0.5 s, and its conversion increases with temperature. At a
temperature of 1173 K, 80% conversion of CHF3 was achieved.
The products of CHF3 pyrolysis detected by GC–MS analysis
include C2F4, C3F6 and trace amounts of C2HF3 at higher
temperatures. At temperatures above 1073 K, trace amounts of
carbonaceous material were formed on the inner surface of the
reactor. As a consequence, the carbon balance drops from 98% at
973 K to around 80% at 1173 K. Fig. 2(b) and (c) shows the
formation rate of major products as a function of temperature. The
production of C2F4 increases with temperature until 1133 K, at this
temperature, a maximum formation rate is achieved before it
decreases at temperatures higher than 1133 K. However, C3F6
formation commences at 1073 K and increases monotonically with
Fig. 1. (a) Apparent 1st order behavior for the pyrolysis of CHF3 over the
temperature range 973–1053 K. X represents fractional conversion of CHF3. (b)
Arrhenius plot for first-order reaction rate constant for the pyrolysis of CHF3.
At pressures between 0.29 and 28.5 bar, Modica et al. found that
the decomposition reaction was first order in CHF3 concentration
[20]. For the lowest pressure (0.29 bar), the reaction could also be
interpreted by a second-order mechanism. Unfortunately, the
calculated results did not reproduce the observed pressure
dependence of the rates for the decomposition of fluoroform [23].
To explain the second-order mechanism, the following reac-
tions were suggested [19]:
ꢄ
CHF3 þ M @ M þ CHF3
(R1)
(R2)
ꢄ
CHF3 ! CF2 þ HF
Table 1
Comparison of kinetic data of CHF3 decomposition obtained in this study with referencesa.
T (K)
Pressure (Pa)
A (sꢁ1 or cm3 molꢁ1 sꢁ1
)
n
E (kJ molꢁ1
)
Reaction order
Ref.
1500–2000
1020–1320
1600–2200
1200–1600
1200–1600
600–2200
973–1073
1150–1570
1600–2200
600–2200
4.93 ꢀ 103–4.93 ꢀ 105
1.29 ꢀ 1014
2.75 ꢀ 1013
7.03 ꢀ 1011
1.26 ꢀ 1012
1.00 ꢀ 1014
9.44 ꢀ 1013
5.2 ꢀ 1013
302
289
244
264
249
294
295
244
244
269
1
1
1
1
1
1
1
2
2
2
[22]
[24]
2.93 ꢀ 104–2.93 ꢀ 106
3.07–4.27 ꢀ 105
[20]
[18]
9.12 ꢀ 104–1.89 ꢀ 106
[19]
[23]
1.01 ꢀ 105
This study
[21]
1.52–2.64 ꢀ 105
2.93 ꢀ 104–2.93 ꢀ 106
2.16 ꢀ 10ꢁ8
2.01 ꢀ 10ꢁ3
1.18 ꢀ 10ꢁ5
ꢁ5.75
[20]
[23]
The rate coefficients of the forward reaction is k = ATn exp(ꢁE/RT), where A is in pre-exponential factor, E is activation energy and R is the ideal gas constant.
a

754
W. Han et al. / Journal of Fluorine Chemistry 131 (2010) 751–760
temperature. At temperatures above 1133 K, its formation rate
exceeds that of C2F4.
et al.’s parameters [22] as shown in Table 1. Similarly to the
conclusions by Biordi et al. [25], these rate parameters appear to be
too high for the pyrolysis of CHF3 compared with experimental
results. The kinetics parameters derived from our experiments,
together with other reported studies, were added to NIST
mechanism to produce a modified mechanism (see Table 2).
The pyrolysis of CHF3 was modelled using the NIST mechanism
and a comparison of experimental data and predicted reactions is
shown in Fig. 2. CHF3 conversion levels agree reasonably with NIST
predictions, although the levels observed experimentally are
slightly higher than those predicted. The NIST mechanism predicts
C2F4 as the sole carbon containing product species and the
predicted rate of formation is consistent with the experimental
results at temperatures below 1093 K. However, it is over-
predicted significantly at temperatures above 1093 K and another
major product, C3F6 which is detected experimentally is not
included as a reaction product in this mechanism. In order to
improve the model’s consistency with experimental data, the NIST
mechanism was modified by incorporating relevant kinetic data
from the open literature. Initially, the kinetics parameters (A factor
and activation energy) of reaction (R2) was replaced by Schug
CF2 þ CF2: CF2 ! C3F6
CF2: CF2 þ CF2: CF2 ! C3F6 þ CF2
c-C3F6 ! CF2: CF2 þ CF2
c-C3F6 ! C3F6
(R10)
(R11)
(R12)
(R13)
In the CHF3 pyrolysis mechanism, the initial reaction step
involves the dehydrofluorination of CHF3, resulting the formation
of HF and CF2 di-radical. Reactions (R3a) and (R3b) are not included
in the mechanism since they are not likely to be responsible for the
consumption of CHF3 as discussed early. One clear deficiency of the
existing mechanism is the absence of reaction pathways which
lead to the formation of C3F6, and as such reactions (R10)–(R13) are
introduced and are part of the modified mechanism. C3F6 also can
be formed via reactions (R14)–(R17), but we suggest these are not
primary pathways as the A factor for the formation of C4F8 ((R14)
and (R15)) is as low as 1010 cm3 molꢁ1 sꢁ1 [27]. In addition, C4F8
was not detected under the conditions studied, suggesting it is not
produced to a significant extent. It is noted that a high activation
barrier was reported by Yu et al. [28] for the transformation of c-
C3F6 to C3F6 (R13). As no c-C3F6 was detected in the present study,
it seems that this reaction does not play a major role in the
formation of C3F6. However, with similar activation energy
(210 kJ molꢁ1) and A factor (1015 ꢁ1), Moon et al. found that this
s
mechanism can predict the trends of C3F6 production and C2F4, as
well for the pyrolysis of CHF3 at 1173 K although the formation rate
of C3F6 is slightly over-predicted [7]. Similarly to the present study,
the intermediate, almost no c-C3F6 was also detected during their
work.
2C2F4 ! C4F8
(R14)
(R15)
(R16)
(R17)
2C2F4 ! c-C4F8
C4F8 ! C3F6 þ CF2
c-C4F8 ! C3F6 þ CF2
Combining the reactions presented in Table 2 with NIST dataset,
results in the development of a modified pyrolysis mechanism.
This mechanism subsequently served to model the reaction
system, the results of which are shown in Fig. 2. The conversion
of CHF3 and C3F6 formation rates predicted by modified mechan-
ism are in good agreement with the experimental values. Although
the prediction of C2F4 formation is improved remarkably compared
with NIST mechanism, a slight over-prediction remains at high
reaction temperatures. Most probably, some by-products formed
at high temperatures, which are not accounted for in the model, are
responsible for this difference.
4.2. Reaction of CHF3 with CH4
Following the study of the pyrolysis of CHF3, the investigation
focused on the reaction of CHF3 with CH4. Our previous studies
have discovered that CHF3 can be converted to vinyl difluoride,
CH255CF2, through reaction with CH4 [8,9]. In these studies, the
reactions involving the formation of a major by-product, C2F4 and
the target product, CH255CF2 were explored, and conditions which
maximized CH255CF2 yield were assessed. To facilitate our under-
Fig. 2. Conversion of CHF3 (a), rate of formation of C2F4 (b) and rate of formation of
C3F6 (c) as a function of temperature during the pyrolysis of CHF3.

W. Han et al. / Journal of Fluorine Chemistry 131 (2010) 751–760
755
Table 2
Modified and new reaction steps considered in modelling of CHF3 pyrolysisa. For the purpose of brevity, reactions taken directly from NIST HFC mechanism are not shown in
this table.
No.
Reaction
A (sꢁ1 or cm3 molꢁ1 sꢁ1
3.4 ꢀ 1030
)
n
E, (kJ molꢁ1
)
Ref.
CHF3 + M ! CF2 + HF + M
ꢁ4.0
288.7
NIST
Replaced by
R2
CHF3 ! CF2 + HF
5.2 ꢀ 1013
1.6 ꢀ 1011
1.0 ꢀ 1012
1.8 ꢀ 1013
6.8 ꢀ 1014
0
0
0
0
0
295
This study
[27]
R10
CF2 + CF2:CF2 ! C3F6
CF2:CF2 + CF2:CF2 ! C3F6 + CF2
c-C3F6 ! CF2:CF2 + CF2
c-C3F6 ! C3F6
77
R11
125.5
182.0
268.8
[47]
R12
[48]
R13
[28]
a
The rate coefficients of the forward reaction is k = ATn exp(ꢁE/RT), where A is in pre-exponential factor, E is activation energy and R is the ideal gas constant.
standing of this reaction and improve the yield of CH255CF2, we
attempt to elucidate the mechanism in more detail and discuss the
pathways which lead to major and minor products.
tures below 1093 K. With further increase of temperature to
1173 K, C, H and F balances drop to 76%, 85% and 71% respectively.
After reaction, soot and a white solid deposit are observed on the
inner surface of reactor and in the alkaline scrubber. We suggest
that the coke and polymer (probably poly-VDF or PTFE) are
responsible for the mass losses, especially at high temperatures.
Increasing the reaction residence time from 0.1 to 0.7 s,
increases the rate of formation of target product, CH255CF2 from
1.5 to almost 3 mmol hꢁ1 at 1173 K as shown in Fig. S1. As the
residence time increases, the rate of formation of C2H2 also shows
significant increase under the conditions studied. Formation of
other minor products, such as C2H3F, C2HF3 and C3F6 are affected
by the change of residence time to a minor extent. It is noted that
the rate of C2F4 formation declines significantly as the residence
time increases from 0.1 to 0.7 s.
In order to improve the yield of CH255CF2, it is important to
explore the mechanism involving its formation and less desirable
reaction by-products. The mechanism of CHF3 conversion is based
on the experimental results of CHF3 pyrolysis, although the
reactions involved in the activation and initial decomposition of
CH4 remain unclear, as there should be no gas-phase decomposi-
tion of CH4 even at 1173 K.
4.2.1. Experimental results
The gas-phase reaction of CHF3 with equimolar CH4 (10% CHF3
and 10% CH4 with N2 balance) was carried out in an alumina tube
reactor at temperature of 873–1173 K, residence time of 0.5 s and
atmospheric pressure. The major products for the reaction of CHF3
with CH4 are CH255CF2, C2F4 and HF. Minor products include C3F6,
CH2F2, C2H3F, C2HF3, C2H6, C2H2 and CHF2CHF2. Trace amounts of
CF3CH55CF2, CH255CFCF3, C4H2F2, C3H8 and CHF55CHF were also
detected by GC–MS. The conversion of CHF3 and CH4 as a function
of temperature is shown in Fig. 3. As expected, the conversion of
CH4 and CHF3 increases with temperature, although CH4 conver-
sion level is always lower than that of CHF3 under the conditions
studied. Conversion of CHF3 commences at 973 K while CH4
conversion was observed at higher temperatures. Fig. 4 shows the
formation rate of major products, CH255CF2 and C2F4, at various
temperatures. It can be seen that once conversion of CH4 is
observed, the target product, CH255CF2 is initiated and similarly to
the conversion of CHF3 and CH4, its selectivity increases
significantly with temperature. Concomitant with CH255CF2
formation, C2F4 is also formed, although a maximum formation
rate of C2F4 is at around 1023 K, above which its rate starts to
decline rapidly. The rate of formation of minor products is
presented in Fig. 5. These products follow a similar rate trend to
C2F4, achieving the highest yield at about 1150 K, except for CH2F2
and C2H2 whose formation rates increase with temperature
monotonically.
4.2.2. Chemical modelling and mechanistic analysis
In the absence of CH4, the conversion of CHF3 amounts to
around 15% at 1073 K as shown in Fig. 2(a). However, with the
introduction of equimolar amounts of CH4 and CHF3, the
conversion level drops slightly to 12%, which is close to conversion
levels observed for CHF3 pyrolysis. In both pyrolysis and in the
presence of CH4, the conversion of CHF3 commences at 973 K.
Clearly, the presence of CH4 does not facilitate the conversion of
CHF3 which suggests that there is a common primary CHF3
Mass balances for C, H and F elements are illustrated in Table S1.
Generally, balances of higher than 95% are achieved at tempera-
Fig. 3. Conversion of CHF3 and CH4 as a function of temperature during the reaction
Fig. 4. Formation rate of major products as a function of temperature during
of CHF3 with CH4.
reaction of CHF3 with CH4.

756
W. Han et al. / Journal of Fluorine Chemistry 131 (2010) 751–760
Fig. 5. Formation rate of minor products as a function of temperature during reaction of CHF3 with CH4.
decomposition pathway for both the processes. Similar results
were found for the thermal decomposition of CHF3 in the presence
of He or H2 [29]. We conclude that CHF3 is inert to attacks by H2, H
or CH4. During pyrolysis of CHF3 and in the reaction with CH4, the
initial reaction steps involve the decomposition of CHF3 and
subsequent formation of HF and the CF2 radical. CH4 is unlikely to
be involved in the initial steps since no significant evidence was
found for its decomposition during CH4 pyrolysis even at 1173 K
[9].
Comparison of the experimental results with model predictions
is shown in Figs. 3–5, including the conversion and rate of
formation of major and minor products. As illustrated in Fig. 3,
although the GRI-NIST model can predict the conversion of CH4
and CHF3 very well, it was found that there are problems
associated with the rate constants of the reactions involved in
CH4 activation [8]. In the GRI-NIST mechanism, reaction pathway
analysis identifies the reverse reaction of (R18) and (R20) and
forward reaction (R19) as these responsible for CH4 activation.
CF3 þ CHF2 ! CF2 þ CHF3
CF3 þ CH4 ! CH3 þ CHF3
CH3 þ CHF2 ! CH4 þ CF2
(R18)
(R19)
(R20)
In contrast, Yu et al. suggests that these reactions are not likely
to be responsible for CH4 activation, since the A factor of
8 ꢀ 1014 cm3 molꢁ1 sꢁ1 is considered to be too high for the reverse
reaction of (R18) [8]. In fact, the singlet CF2 is very non-reactive
since the closed-shell singlet CF2 (1A1) is strongly stabilized by pp-
back donation [30]. However, the metastable triplet CF2 (3B1),
having a rather long lifetime of about 1 s, is believed to be much
more reactive and an A factor of 1.2 ꢀ 1013 cm3 molꢁ1 sꢁ1 was
estimated for the reaction with CH4 [31]. Because the triplet lies
238.1 kJ molꢁ1 above the ground-state singlet, the ratio of triplet to

W. Han et al. / Journal of Fluorine Chemistry 131 (2010) 751–760
757
singlet ground-state populations is only 8 ꢀ 10ꢁ11, hence the
reactions of the triplet can be neglected [8].
By means of ab initio calculations, we failed to locate a transition
structure for the reverse reaction (R20). Instead, it was found that
CF2 can insert into C–H via reaction (R21) with activation energy of
163 kJ/mol. The exponential factor was estimated to be
In the previous [8] and present studies, trace amounts of H2 were
detected, although accurate determination of the amount of H2
produced was not possiblewiththe present analysistrain. Excluding
H2, over 95% of hydrogen is balanced, which supports the assertion
that the rate of formation of H2 is low. However, it has been
suggested that a relatively low concentration of Hꢅ may lead to the
activation of CH4 via a series of chain reactions ((R27)–(R31)) [33].
Initiation:
1.5 ꢀ 1011 cm3 molꢁ1 sꢁ1
CF2 þ CH4 ! CH3CF2H
.
(R21)
(R22)
CH4ðsÞ ! CH3 þ H
Propagation:
(R27)
CH3CF2H ! CHF2 þ CH3
With even higher activation energy (401 kJ molꢁ1), CH3CF2H
can further produce CH3 and CHF2 radicals through reaction (R22).
The exponential factor of this reaction is around
1 ꢀ 1017 cm3 molꢁ1 sꢁ1 based on the analogous decomposition
reaction of C2H6.
H þ CH4 ! CH3 þ H2
(R28)
(R29)
(R30)
(R31)
CH4 þ CH3 ! C2H6 þ H
CH3 þ C2H6 ! C2H5 þ H
C2H5 ! C2H4 þ H
CH3 þ CF2 ! CH2¼CF2 þ H
(R23)
Once CH3 is formed, the major product, CH255CF2 can be
produced via reaction (R23) with CF2, which is derived from the
elimination of HF from CHF3. This reaction has been studied using
quantum chemical theory and it is found that this reaction leads to
the formation of CH255CF2 with an activation energy of almost zero
[32]. With this discussion in mind, new reaction steps including
(R21)–(R23) were introduced into the GRI-NIST mechanism,
replacing (R18)–(R20). However, modelling results show that this
new mechanism does not correctly predict the conversion of CH4.
Apparently, reactions (R21) and (R22) are not the primary
pathways responsible for the activation of CH4 and formation
CH3 because of their high-energy barriers. There must be other
reactions playing a major role in the activation of CH4.
Toexaminethishypothesis,thepyrolysisofCH4 wasinvestigated
both experimentally and computationally. The results are shown in
Table 3, together with the modelling results based on the existing
GRI mechanism. Indeed, the conversion of CH4 and subsequent
formation of C2H6, although low, was observed with our alumina
reactor. The existing GRI mechanism was used to model this
reaction. Virtually no reaction was predicted under any of the
conditions studied. This suggests that a small amount of CH4 may be
activated on the surface of reactor, at least at high temperatures.
During the reaction of CH4 with CHF3, as a result of the production of
HF, the reactor surface is likely to be fluorinated [34]. Hence, 0.2 g
AlF3 (having a bulk volume of 0.35 cm3 and a surface area of
480 cm2) was packed into the reactor to simulate this fluorinated
reactor surface. As shown in Table 3, AlF3 enhances the rate of
decomposition of CH4 significantly. Consistent with these observa-
tions, it has been noted that methane can be activated, or its
conversion can be improved, in the presence of Brønsted and/or
Lewis acid sites [35–38]. AlF3 is considered a strong solid Lewis acid
[39] and it may initiate chain reactions ((R27)–(R31)). Although CH4
is stable even at 1173 K, low conversion levels still can be achieved
on the surface of reactor. Based on these results, we introduce into
the existing mechanisms a simplified reaction step(R27) to simulate
thesesurfacereactions.Numerousresultshavebeenreportedforthe
kinetics of gas-phase reaction of (R27) [40,41], for which the
estimated activation energy varies from 270 to 490 kJ molꢁ1. Based
on the observed effect of surface reaction on methane activation, an
activation energy of 270 kJ molꢁ1 appears to be too high. Improved
agreement with experimental data can be obtained by decreasing
the activation energy and exponential factor to 234 kJ molꢁ1 and
It was speculated that there are reactions occurring on the
surface of the reactor (a-Al2O3) and CH4 activation takes place as a
consequence of surface reactions [8]. In order to model the reaction
system, a reaction step (R24) was included to mimic the surface
reaction while CH4 was thought to be activated by surface fluorine
radicals. However, this hypothesis seems to conflict with experi-
mental results when it was found that after packing Al2O3 or AlF3
chips into the reactor, no enhanced conversion of CH4 was
observed. Furthermore, if this reaction takes place to a significant
extent, the formation of CF4 and C2F6 via reactions (R7), (R25) and
(R26) during pyrolysis of CHF3 is expected. However, even trace
amounts of CF4 and C2F6 were not detected in the present study. In
addition, for the reaction of CHF3 with CH4, CH3F will be the key
intermediate which is also absent from the product profile.
CF3 þ CF3 ! C2F6
CF2 ! C þ F þ F
CF2 þ F ! CF3
CF3 þ F ! CF4
Table 3
(R7)
(R24)
(R25)
(R26)
1.09 ꢀ 1010
s
ꢁ1. After introducing this reaction into GRI mechanism,
closer prediction can be obtained compared with experimental
results as shown in Table 3.
Combining the other reaction steps outlined in Table 4 with the
GRI-NIST mechanism, we introduce our new modified reaction
Experimental and modelling results of pyrolysis of CH4 as a function of temperature in an alumina tube reactora.
b
T (K)
CH4 Pyrolysis
GRI-Mech Prediction
Modified model
CH4 Pyrolysis over AlF3
Conversion
(%)
C2H6
(mmol hꢁ1
Conversion
(%)
C2H6
(mmol hꢁ1
Conversion
(%)
C2H6
(mmol hꢁ1
Conversion
(%)
C2H6
(mmol hꢁ1
C2H4
(mmol hꢁ1
)
)
)
)
)
973
1023
1073
1123
1173
0.049
0.024
0.018
0.14
0
0.002
0.004
0.008
0.02
0.002
0.018
0.098
0.46
0.00014
0.0016
0.009
0.04
0.11
0.26
0.68
0.37
1.1
0.0036
0.0061
0.01
0
0.0026
0.0018
0.004
0.013
9 ꢀ 10ꢁ7
0
8.6 ꢀ 10ꢁ6
0.002
0.005
0.009
7 ꢀ 10ꢁ5
0.011
0.027
0.31
0.04
0.0005
0.83
0.011
a
At pressure of 1.01 bar and residence of 0.5 s.
b
During experiment, 0.2 g AlF3 was packed into the reactor. Other reaction conditions remained unchanged.

758
W. Han et al. / Journal of Fluorine Chemistry 131 (2010) 751–760
Table 4
Modified and new reaction steps considered in modelling of reaction of CHF3 with CH4a. For the purpose of brevity, reactions taken directly from NIST HFC mechanism are not
shown in this table.
No.
Reaction
A (sꢁ1 or cm3 molꢁ1 sꢁ1
3.4 ꢀ 1030
)
n
E (kJ molꢁ1
)
Ref.
CHF3 + M ! CF2 + HF+ M
ꢁ4.0
288.7
NIST
Replaced by
R2
CHF3 ! CF2 + HF
5.2 ꢀ 1013
1.6 ꢀ 1011
1.0 ꢀ 1012
1.8 ꢀ 1013
6.8 ꢀ 1014
3.0 ꢀ 1013
0
0
0
0
0
0
295
This study
[27]
R10
CF2 + CF2:CF2 ! C3F6
CF2:CF2 + CF2:CF2 ! C3F6 + CF2
c-C3F6 ! CF2:CF2 + CF2
c-C3F6 ! C3F6
77
R11
125.5
182.0
268.8
3.4
[47]
R12
[48]
R13
[28]
R20
CHF2 + CH3 ! CF2 + CH4
Replaced by
ꢁR20
R21
CH4 + CF2 ! CH3 + CHF2
CH4 + CF2 ! CH3CF2H
CH3CF2H ! CH3 + CHF2
CH3 + CF2 ! CH2CF2 + H
1.0 ꢀ 1013
1.5 ꢀ 1011
1.0 ꢀ 1017
6.0 ꢀ1012
0
0
0
0
159.5
163
[8]
This study
This study
R22
401
R23
14.6
Change to
2.1 ꢀ 1013
1.09 ꢀ 1010
2.0 ꢀ 1012
ꢁ0.2
0
234
14.6
[32]
R27
R34
CH4 ! H + CH3
0
0
This study
CHF2 + CF2 ! CF2:CF2 +H
Replaced by
ꢁR34
CF2:CF2 + H ! CHF2 + CF2
8.4 ꢀ 108
1.5
19.2
This study
The rate coefficients of the forward reaction is k = ATnexp(ꢁE/RT), where A is in pre-exponential factor, E is activation energy and R is the ideal gas constant.
a
mechanism. The reaction parameter for CHF3 decomposition is
obtained from our pyrolysis study. Relatively good agreement
between model predictions and experimental results are achieved,
as shown in Figs. 3–5, although among the products, C2HF3 and
C3F6 are significantly over-predicted by the modelling. We suggest
this is due to the absence of reactions steps which lead to the
decomposition of these two species. For example, significant
amounts of soot were observed during the experiments, while only
trace amounts of carbon were predicted from modelling.
With this modified mechanism, sensitivity analysis for the
concentration of CH3 was performed and the results are shown in
Fig. 6. Sensitivity analysis is often used to describe the dependence
of rate of formation of a certain species on the reactions involved in
the mechanism. By comparing sensitivity coefficients, the rate-
determining and other most important reactions for formation of
different species can be identified and analyzed quantitatively. A
more detailed example can be seen in Reference [42].
this reaction to CH3 concentration is extremely low. This indicates
that the primary role for reaction (R27) is that it acts as an initiator
for the conversion of CH4.
Initiation:
CH4ðsÞ ! ꢅ CH3 þ Hꢅ
Propagation:
(R27)
H þ CF2: CF2 ! CHF2CF2
(R32)
(R33)
H þ CF2: CF2 ! CHFCF2 þ F
CHF2 þ CF2 ! H þ CF2: CF2
CHF2CF2 þ CH4 ! CHF2CHF2 þ CH3
F þ CH4 ! CH3 þ HF
(R34)
(R35)
(R36)
(R37)
(R23)
Clearly, the sensitivity analysis further confirms our suggestion
of the major pathways for activation of CH4. Although almost no
CH4 conversion was observed without reaction (R27), sensitivity of
H þ CH4 ! CH3 þ H2
CH3 þ CF2 ! CH2¼CF2 þ H
Fig. 6. Sensitivity analysis of reactions to the concentration of CH3 radical at temperature 1113 K, pressure of 1.01 bar and residence time of 0.5 s.

W. Han et al. / Journal of Fluorine Chemistry 131 (2010) 751–760
759
As shown in Fig. 6, CH3 consumption is dominated by reaction
5. Conclusions
(R23), forming the target product, CH255CF2. In reaction (R23), CF2
is formed via reaction (R2) which is the major channel for the
decomposition of CHF3 (see Fig. S2).
The pyrolysis of CHF3 and reaction of CHF3 with CH4 have been
investigated experimentally and computationally. It was found
that CHF3 pyrolysis commences around at 973 K at residence time
of 0.5 s, and its conversion increases with temperature. The
products identified by GC–MS include C2F4, C3F6 and trace amounts
For the NIST-GRI mechanism, the major reaction step leading to
the formation of C2HF3 involves the coupling of CHF2 and CF2 via
reaction (R38). Although no literature references for this reaction
have been found, kinetic data in the NIST-GRI mechanism is
consistent with the results of similar reactions, which have been
investigated both experimentally and computationally [43]. We
suggest that the final estimation of the concentration of C2HF3 is
over-predicted because of the relatively high concentration of
CHF2 predicted in the model.
of C2HF3. The overall rate of CHF3 decomposition can be expressed
ꢁ1
as 5:2 ꢀ 1013 ½sꢁ1ꢂ eꢁ295½kJ mol ꢂ=RT . The NIST mechanism predicts
C2F4 as the only carbon containing species, with its formation rate
over-predicted significantly at temperatures above 1093 K. An-
other major reaction product, C3F6 is not included in the
mechanism. A modified mechanism was developed which can
reproduce experimental data reasonably well.
CHF2 þ CF2 ! CHF : CF2 þ F
(R38)
For the reaction of CHF3 with CH4, the major products are
CH255CF2, C2F4 and HF. Minor products include C3F6, CH2F2,
C2H3F, C2HF3, C2H6, C2H2 and CHF2CHF2. Trace amounts of
CF3CH55CF2, CH255CFCF3, C4H2F2, C3H8 and CHF55CHF were also
detected by GC–MS. Predictions based on the GRI-NIST
mechanism show good agreement between experiments and
modelling for the conversion of CHF3 and CH4, but significant
discrepancies were observed for the selectivity of some reaction
products. Modifications to the existing GRI-NIST mechanism are
suggested and when incorporated in the model, significantly
improve agreement between experiments and modelling. Based
on mechanistic pathway analysis, the initial step in the
decomposition of CHF3 includes the formation of CF2 radical
and HF. This reaction dominates the pyrolysis of CHF3 and
reaction of CHF3 with CH4. Trace amounts of CH4 decompose on
the surface of reactor, producing H radical. It is proposed that
CH4 is activated by a series of chain reactions which are initiated
by this small amount of H radicals.
Reaction pathway analysis indicates that the main pathway to
CHF2 formation is via the reverse reaction of (R34). Its second-
order rate constant is estimated to be greater than
6 ꢀ 10ꢁ8 cm3 molꢁ1 sꢁ1 at 1093 K, based on the modelling data,
although much lower values were reported [44–46] for this
reaction based on the experimental observation. To improve the
prediction of C2HF3, (R34) was replaced by (ꢁR34) in the modified
mechanism. The activation energy is assumed to be 19 kJ molꢁ1
,
which is equal to the standard enthalpy changes of (ꢁR34), which
is lower than the activation energy barrier.
There are a number of uncertainties in elucidating the pathways
for formation of C3F6, which is generally thought to be formed via
reaction of C2F4 with CF2 [27,28]. The reaction between CF2 and
C2F4 was studied using the meta hybrid density functional theory
method of BB1K, which showed that the addition of CF2 to C2F4
invariably leads to the formation of c-C3F6, although no transition
state leading directly to C3F6 was found [28]. In the present
experiments, no c-C3F6 was detected, which suggests the NIST-GRI
mechanism may not include all pathways leading to the formation
of C3F6. As a consequence, significant discrepancies were found
between the prediction and experimental results of C3F6. In order
to model the formation of C3F6, reactions (R10)–(R14) are included
in the modified mechanism, which result in reasonable agreement
with experiments during the pyrolysis of CHF3.
Acknowledgements
The Australian Research Council is gratefully acknowledged for
financial support for this project. W.H. is indebted to the
Department of Education, Science and Training (DEST) of the
Australian Government and the University of Newcastle, Australia
for postgraduate scholarships.
Based on the suggested mechanism, major channels to other
products can be obtained.
Channel to C2H6:
Appendix A. Supplementary data
2CH3 ! C2H6
(R39)
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jfluchem.2010.03.012.
Channels to CH2F2:
CHF2 þ CH4 ! CH2F2 þ CH3
(R40)
(R41)
(R42)
References
CHF3 þ CHF2 ! CH2F2 þ CF3
[1] Production and Consumption of Ozone Depleting Substances Under the Montreal
Protocol, UNEP, 2008 Available at http://ozone.unep.org/Data_Reporting/Data
_Access/.
[2] D.W.F. Guus, J.M. Velders, J.S. Daniel, M. McFarland, S.O. Andersen, Proc. Nat. Acad.
Sci. 106 (2009) 10949–10954.
2CHF2 ! CH2F2 þ CF2
Channels to C2H3F:
CH4 þ CF2 ! CH2: CHF þ HF
(R43)
(R44)
(R45)
(R46)
[3] D.E. Oram, W.T. Sturges, S.A. Penkett, A. McCulloch, P.J. Fraser, Geophys. Res. Lett.
25 (1998) 35–38.
[4] AM0001: Incineration of HFC 23 Waste Streams, UNFCCC, 2006http://cdm.unfccc.
int/methodologies/index.html.
[5] W.B. Feaver, J.A. Rossin, Catal. Today 54 (1999) 13–22.
[6] H. Onoda, T. Ohta, J. Tamaki, K. Kojima, Appl. Catal. A 288 (2005) 98–103.
[7] D.J. Moon, M.J. Chung, H. Kim, Y.S. Kwon, B.S. Ahn, Ind. Eng. Chem. Res. 41 (2002)
2895–2902.
[8] H. Yu, E.M. Kennedy, J.C. Mackie, B.Z. Dlugogorski, Environ. Sci. Technol. 40 (2006)
5778–5785.
CH4 þ CF ! CH2: CHF þ H
CH3 þ CHF2 ! CH2: CHF þ HF
CH2¼CF2 þ H ! CH2: CHF þ F
Channels for CHF2CHF2:
2CHF2 ! CHF2CHF2
[9] W.F. Han, H. Yu, E.M. Kennedy, J.C. Mackie, B.Z. Dlugogorski, Environ. Sci. Technol.
42 (2008) 5795–5799.
(R47)
(R48)
(R35)
[10] A.L. Moore, Fluoroelastomers Handbook: the Definitive User’s Guide and
Databook, William Andrew Publishing, Norwich, 2005.
[11] M. Visca, M.S. Kelly, B.L. Kent, EP 1679352 (2007).
[12] H. Yu, E.M. Kennedy, M.A. Uddin, A.A. Adesina, B.Z. Dlugogorski, Catal. Today 97
(2004) 205–215.
CHF2CF2 þ H ! CHF2CHF2
CHF2CF2 þ CH4 ! CHF2CHF2 þ CH3

760
W. Han et al. / Journal of Fluorine Chemistry 131 (2010) 751–760
[13] K. Li, E.M. Kennedy, B.Z. Dlugogorski, Ind. Eng. Chem. Res 38 (1999) 3345–3352.
[14] M.J. Height, E.M. Kennedy, B.Z. Dlugogorski, J. Chromatogr. A 841 (1999) 187–195.
[15] Softpredict, Cosilab Collection, Version 2.1.0 ed., Rolexo-Softpredict-Cosilab
GmbH & Co KG, Bad Zwischenahn, Germany, 2007, www.Softpredict.com.
[16] A. Manion, R.E. Huie, R.D. Levin, D.R. Burgess Jr., V.L. Orkin, W. Tsang, W.S.
McGivern, J.W. Hudgens, V.D. Knyazev, D.B. Atkinson, E. Chai, A.M. Tereza, C.-
Y. Lin, T.C. Allison, W.G. Mallard, F. Westley, J.T. Herron, R.F. Hampson, D.H.
Frizzell, NIST Chemical Kinetics Database, NIST Standard Reference Database 17,
Version 7.0 (Web Version) Release 1.4.3, Data Version 2008.12, National Institute
of Standards and Technology, Gaithersburg, MD, 2008, http://kinetics.nist.gov/.
[17] G.P. Smith, D.M. Golden, M. Frenklach, N.W. Moriarty, B. Eiteneer, M. Goldenberg,
C.T. Bowman, R.K. Hanson, S. Song, W.C. Gardiner, V.V. Lissianski Jr., Z. Qin, GRI-
Mech 3.0, 2008 http://www.me.berkeley.edu/gri_mech/.
[29] R. Romelaer, V. Kruger, J.M. Baker, W.R. Dolbier, J. Am. Chem. Soc. 123 (2001)
6767–6772.
[30] S. Koda, J. Phys. Chem. 83 (1979) 2065–2073.
[31] X.J. Hou, T.L. Nguyen, S.A. Carl, J. Peeters, M.T. Nguyen, Chem. Phys. Lett. 402
(2005) 460–467.
[32] H. Yu, J.C. Mackie, E.M. Kennedy, B.Z. Dlugogorski, Ind. Eng. Chem. Res. 45 (2006)
3758–3762.
[33] C.H. Bamford, C.F.H. Tipper, in: C.F.H.T.C.H. Bamford (Ed.), Decomposition and
Isomerisation of Organic Compounds, Elsevier, Amsterdam, 1972, pp. 36–64.
[34] M. Bankhead, G.W. Watson, G.J. Hutchings, J. Scott, D.J. Willock, Appl. Catal. A 200
(2000) 263–274.
[35] T. Jin, T. Yamaguchi, K. Tanabe, J. Phys. Chem. 90 (2002) 4794–4796.
[36] W. Hua, A. Goeppert, J. Sommer, Appl. Catal. A 219 (2001) 201–207.
[37] F.T.T. Ng, C.R. Rourke, J. Lynn, J.S. Kevin, E.C. Sanford, Studies in Surface Science
and Catalysis, Elsevier, 1992, pp. 91–97.
[18] E. Tschuikow-Roux, J.E. Marte, J. Chem. Phys. 42 (1965) 2049–2056.
[19] E. Tschuikow-Roux, J. Chem. Phys. 42 (1965) 3639–3642.
[20] A.P. Modica, J.E. LaGraff, J. Chem. Phys. 44 (1966) 3375–3379.
[21] Y. Hidaka, T. Nakamura, H. Kawano, Chem. Phys. Lett. 187 (1991) 40–44.
[22] K.P. Schug, H.G. Wagner, F. Zabel, Ber. Bunsen-Ges. Phys. Chem. 83 (1979) 167–
175.
[38] F.R. Chen, G. Coudurier, J.F. Joly, J.C. Vedrine, J. Catal. 143 (1993) 616–626.
[39] G. Eltanany, S. Rudiger, E. Kemnitz, J. Mater. Chem. 18 (2008) 2268–2275.
[40] W.C. Gardiner Jr., J.H. Owen, T.C. Clark, J.E. Dove, S.H. Bauer, J.A. Miller, W.J.
McLean, Symp. (Int.) Combust. 15 (1975) 857–868.
[23] D.W. Placzek, B.S. Rabinovitch, G.Z. Whitten, E. Tschuikow-Roux, J. Chem. Phys. 43
(1965) 4071–4080.
[41] J.H. Kiefer, S.S. Kumaran, J. Phys. Chem. 97 (1993) 414–420.
[42] M. Forsth, Combust. Flame 130 (2002) 241–260.
[24] S.F. Politanskii, V.U. Shevchuk, Kinet. Catal. 9 (1968) 496–503.
[25] J.C. Biordi, C.P. Lazzara, J.F. Papp, J. Phys. Chem. 82 (1978) 125–132.
[26] F. Battin-Leclerc, A.P. Smith, G.D. Hayman, T.P. Murrells, J. Chem. Soc. Faraday
Trans. 92 (1996) 3305–3313.
[43] Y. Saso, D.L. Zhu, H. Wang, C.K. Law, N. Saito, Combust. Flame 114 (1998) 457–468.
[44] K. Sugawara, K. Okasaki, S. Sato, Bull. Chem. Soc. Jpn. 54 (1981) 358–361.
[45] E.B. Gordon, B.I. Ivanov, A.P. Perminov, V.E. Balalaev, Chem. Phys. 35 (1978) 79–
89.
[27] N.N. Buravtsev, A.S. Grigor’ev, Y.A. Kolbanovskii, Kinet. Catal. 30 (1989) 21–30.
[28] H. Yu, E.M. Kennedy, W.H. Ong, J.C. Mackie, W. Han, B.Z. Dlugogorski, Ind. Eng.
Chem. Res. 47 (2008) 2579–2584.
[46] J.P. Kilcoyne, K.R. Jennings, J. Chem. Soc., Faraday Trans. 1 (70) (1974) 379–390.
[47] B. Atkinson, V.A. Atkinson, J. Chem. Soc. (1957) 2086–2094.
[48] S.H. Bauer, S. Javanovic, Int. J. Chem. Kinet. 30 (1998) 171–177.